1. Field of the Invention
The invention concerns a method for making microcavities on a substrate. It can be applied to an electrochemical sensor for measuring the concentration of reactive species with reduced response time and, more particularly, to a sensor made with thin film or thick film type technology on a substrate that is electrically insulating and chemically inert at high temperature. It also concerns a method for making a sensor of this type.
The invention can also be applied, inter alia, to the making of a gas-phase chromatograph and, especially, to the separation column of a chromatograph of this type.
2. Description of the Prior Art
One of the well-known groups of electrochemical sensors works on the principle of the concentration cell and measures the partial pressure of one or more species of the gaseous mixture to be analyzed. This gaseous mixture, which is present in a first compartment and is, for example, an oxygen/inert gas mixture, is separated from a reference medium by the wall of a solid electrolyte, each face of which has an electrode. As is well known, the equations that govern the working of these sensors are: at the electrodes/electrolyte interfaces: ##STR1## the voltage V.sub.E1/E2 which then develops; between the electrodes is given by Nernst's law: ##EQU1## with
R=perfect gas constant=8.314 J.(mole. K).sup.-1
F=Faraday No.=96490 Coulombs
T=absolute temperature in degrees Kelvin
P.sub.1 and P.sub.2 =partial pressures of media 1 and 2 in the compartments 1 and 2.
Thus, knowledge of the temperature and of one of the partial pressures enables the unambiguous determining of the other partial pressure.
Should the mixture be reactive, for example, if it is a mixture of O.sub.2 and CO, and if the electrode is a catalyst of the reaction of these gases, the following reaction occurs: EQU 2CO+O.sub.2 .revreaction.2CO.sub.2 ( 3)
and, finally, if the combustion is total, until reversible thermodynamic equilibrium is achieved, the following relationship is verified: ##EQU2## with K(T) being a coefficient of equilibrium depending on the temperature, and P CO, P O.sub.2, P CO.sub.2, being the partial pressures of carbon dioxide, oxygen and carbonic gas.
In applications concerning the regulation of automobile engines with spark ignition, to determine the partial pressure of oxygen at the exhaust (medium 1 for example), knowing the reference pressure (medium 2 which is generally air) in removing the need to measure or regulate the temperature, use is made of the fact that, if the exhaust gases are brought to thermodynamic equilibrium (end of combustion), the value of the partial pressure of oxygen, as shown in FIG. 1, varies by some fifteen orders of magnitude when the mixture feeding the cylinders passes through the stoichiometric state.
Thus, in the above-described Nernst formula, a voltage leap is observed when the mixture passes the stoichiometric state .DELTA.V=RT/4F log PO.sub.2 rich/PO.sub.2 poor); if the temperature is in the range of 800.degree. C., the term RT/4F is of the order of 50 mV and the .DELTA.V will be greater than 750 mV.
Sensors of this type, called stoichiometrical sensors, generally consist of a glove finger made of stabilized zircon. The external wall, provided with a porous platinum electrode (measuring electrode), is in contact with the gas for which it is sought to analyze the oxygen content and the inner wall, also provided with a platinum electrode (reference electrode), is in contact with a reference gas, generally air. The platinum of the measuring electrode catalyzes the end of combustion of the exhaust gases for example and, in order not to saturate the platinum, it is encapsulated by means of a porous diffusion layer which has the main effect of limiting the flow of gases reaching the catalytic sites of the platinum electrode.
FIG. 2 shows a few typical responses of these glove finger sensors using air as a reference.
However, the making of sensors of this type may take different forms. FIGS. 3 and 4 show examples of methods for making these sensors obtained by successive deposits (thin layers or thick layers) of ceramic and metallic materials on an electrically insulating substrate. According to FIG. 3, there is a known method to make an electrochemical sensor comprising a solid electrolyte EL on a substrate Sb. This electrolyte may be made of zircon, thoria or cerium oxide stabilized by one or more elements belonging to columns II.sub.A and III.sub.B of the periodic classification of elements. It may be made as a thin layer or a thick layer, or it may be massive.
Electrodes E2/P2 and E1/P1 are deposited on the electrolyte EL and on the substrate Sb. The electrodes E1/P1 and E2/P2 are located in one and the same plane. The electrode E1/P1 combines the functions of an electrode and a reference medium. The electrode E1/P1 is further protected from the external environment by an impervious and inert insulating material S1 which coats it. It is possible, for example, to use an association of the type Ni/NiO or Pd/PdO to make this electrode/reference medium. The electrode E2/P2 has two zones and communicates directly with the medium to be analyzed in which there flows the gaseous mixture G through a hole made in the insulating body S1 which also covers it. In the first zone Ct, the electrode is not in contact with the electrolyte EL. The fluid to be analyzed must flow through the zone Ct which takes the place of a catalyst. In this zone, the reactive species of the mixture to be analyzed (for example, in the case of exhaust gases: CO and O.sub.2) are brought to full thermodynamic equilibrium before they have reached the electrochemical cell itself: EQU E2/P2-EL-E1/P1
P2 represents the partial pressure of oxygen after catalysis in the real medium to be analyzed. The catalysis which enables obtaining thermodynamic equilibrium is achieved by the fact that the fluid flows through the catalyst in a direction parallel to the plane of the electrodes. The electrodes are extended outwards by metallic links to which the contacts C1 and C2 may be soldered. These links are made with platinum veneer for example. In one practical embodiment, the metallic links and the electrodes are made so as to form a single part. The substrate Sb may consist of a material (such as corundum) which insulates well at the operating temperature of the device and gives the unit mechanical strength. The face of the substrate 1 opposite the electrochemical cell has a heating resistor RC which enables accelerating the reaction.
The deposits can be made by well known techniques, such as: vacuum deposition (cathode spraying, evaporation), vapor phase deposition, electrochemical deposition or ion implantation or by a combination of two or more of these techniques. For a metal/oxidated metal reference mixture, such as Pd/PdO, the response, in voltage, to a temperature of about 800.degree. C. is shown in FIG. 2 for the corresponding temperature.
The descriptions of sensors thus made will be found in the French patent Nos. 2441 164 and 2 444 272.
FIG. 5 shows another embodiment of a sensor according to the prior art.
This figure repeats the elements illustrated with reference to FIG. 3: the measuring cell E1/P1-El1-E2/P2, deposited in thin or thick layers or massively on a substrate Sb, the catalysis region Ct and the test samples intake region P.sub.es where the interactions with the gaseous mixture to be analyzed take place. In fact, in the example described, these latter two regions consist of an extension of the measuring electrode E2/P2. The output signal VS of the sensor is transmitted to external circuits (not shown) by connections C1 and C2. The two electrodes E1/P1 and E2/P2 should at least be shielded by an impervious and inert insulating jacket S1, made of enamel for example.
According to the sensor of FIG. 5, an additional electrochemical cell is integrated into the sensor and comprises a solid electrolyte El2 inserted between two electrodes E3 and E4. In the embodiment of FIG. 5, and according to the first approach, the second electrode E4 is identified with the extension of the measuring electrode E2. The cell is flush with the surface of the insulating material S1 so as to communicate with a medium containing oxygen. This medium may be the medium Mex in which there flows the gaseous mixture G to be analyzed. The cell E3-El2-E4 is supplied with a control current Ip by means of connections C3 and C4, C4 being identified with C2. The face of the substrate opposite to the electrochemical cell also has a heating resistor RC.
Referring again to the above description, it is immediately seen that the cell E3-El2-E4, working as an ion pump, modifies the oxygen composition of the test sample let the sensor, namely the oxygen composition of the gaseous mixture flowing towards and through the catalysis zone Ct to subsequently reach the measuring cell E2/P2-El1-E1/P1, and does this modification as a function of the amplitude and bias of the current Ip. It follows that this cell produces an output signal VS which flips over, no longer when the stoichiometric state of the mixture G is reached but "before" or "after" said stochiometric state, the lag on either side of the stoichiometric state being defined continuously by the amplitude and bias of the control current Ip. FIG. 6 shows a few typical responses of this type of sensor as a function of the bias current Ip.
A description of a sensor of this type will be found in French patent Nos. 2 494 445 and 2 442 444.
In prior art sensors such as those described above, the gaseous mixture is thus catalyzed, making it possible to obtain a thermodynamic equilibrium, and to do so before measuring the concentration of the species constituting the mixture. This catalysis is performed in the course of a diffusive path in a porous body. This path is laid down by the geometry of the sensor and the different deposits constituting the sensor.
This diffusive path causes a delay in the analysis of the gas. This delay is due to the time taken by the different molecules to move forward between the place where they are incorporated in the porous material (test sample window G) and the place where they will fix the electrochemical potential of the working electrode. This delay in the analysis of the gases can penalize the total response time of the sensor (since this sensor furthermore brings into play phenomena in which the different electrodes are placed in equilibrium). This diffusion time, with a delay .tau., can be evaluated by the formula: .tau.=l.sup.2 /D where l is the length of the diffusive path and D is the coefficient of diffusion of the gaseous species in the porous medium.
If the mean diameter of the pores of the diffusive medium is in the range of the length of the mean free path of the gas molecules at working temperature and pressure (between 10.sup.2 and 10.sup.3 angstroms in an automobile exhaust silencer for example), the diffusion takes places according to a principle called Knudsen's flow according to which the interaction of the molecules takes place preponderantly with the walls of the pores. In this case, the expression of the diffusion coefficient is: EQU D.sub. k=2/3a(8kT/.pi.M).sup.1/2
where is the mean radius of the pores, k is the Boltzmann constant, T is the absolute temperature and M is the molecular mass of the diffusing gaseous species. For example, the coefficient of diffusion of oxygen at 1000.degree. K. and at atmospheric pressure in a porous medium where the mean diameter of the pores is 10.sup.3 angstroms is about 10.sup.-2 cm.sup.2 sec.sup.-1. The diffusion time associated with a diffusion length of about 100 um is about 100 ms in these conditions.
It might be further observed that, in Knudsen's formula, only purely mechanical interactions between the gas and the walls of the pores are taken into account.
Now, when reactive gases such as CO and O.sub.2 are made to diffuse in a porous medium consisting of oxides or metal/oxide (Cermet) mixtures, the impacts of the gas molecules on the walls of the pores are accompanied by a retention time, corresponding to a transient adsorption of the gas molecules at the faults on the surface which are always present in oxide or metal polycrystals. Thus, the molecules move forward by successive adsorption and desorption processes: the diffusion time of the gas molecules in question is thereby considerably increased and may go up to several seconds for a diffusion length always equal to 100 .mu.m.
If, on the other hand, the diffusive medium is such that the mean size of the pores is far greater than the mean free path of the gas molecules in the medium at the pressure and temperature considered, the molecules will interact with one another preponderantly, and the effect of the walls of the pores will tend to disappear. In this case, the diffusion occurs by successive impacts of the diffusing gas molecules on the molecules of the carrier gas. The diffusion coefficient of the gas i in a monomolecular carrier gas j is then written in a simplified way as follows ##EQU3##
where the Mi,j are molecular masses of the gases i and j, T is the absolute temperature, p is the total pressure, 6ij is the diameter of collision and K is a constant. For example, the diffusion coefficient of oxygen at 1000.degree. K. in nitrogen is about 1.6 cm.sup.2.sec.sup.-1. Thus, the diffusion time associated with a diffusive length of about 100 .mu.m is, in these conditions, about 100 .mu.sec, i.e. at least 100 times smaller than in the preceding case considered.
It would therefore seem, in the light of these assessments, that the more the diffusive mode in the porous limiting layer is of a molecular type (the second example described above), the shorter will be the associated diffusion time in the present example, and the shorter will be the response time of the sensor.
The invention therefore concerns a sensor in which the molecular diffusion mode is favored in order to reduce the diffusion time of the gas molecules and, correlatively, in order to reduce the total response time of the sensor.